Intracellular Trafficking of AIP56, an NF-κB-Cleaving Toxin from Photobacterium damselae subsp. piscicida

Abstract

AIP56 (apoptosis-inducing protein of 56 kDa) is a metalloprotease AB toxin secreted by Photobacterium damselae subsp. piscicida that acts by cleaving NF-κB. During infection, AIP56 spreads systemically and depletes phagocytes by postapoptotic secondary necrosis, impairing the host phagocytic defense and contributing to the genesis of infection-associated necrotic lesions. Here we show that mouse bone marrow-derived macrophages (mBMDM) intoxicated by AIP56 undergo NF-κB p65 depletion and apoptosis. Similarly to what was reported for sea bass phagocytes, intoxication of mBMDM involves interaction of AIP56 C-terminal region with cell surface components, suggesting the existence of a conserved receptor. Biochemical approaches and confocal microscopy revealed that AIP56 undergoes clathrin-dependent endocytosis, reaches early endosomes, and follows the recycling pathway. Translocation of AIP56 into the cytosol requires endosome acidification, and an acidic pulse triggers translocation of cell surface-bound AIP56 into the cytosol. Accordingly, at acidic pH, AIP56 becomes more hydrophobic, interacting with artificial lipid bilayer membranes. Altogether, these data indicate that AIP56 is a short-trip toxin that reaches the cytosol using an acidic-pH-dependent mechanism, probably from early endosomes. Usually, for short-trip AB toxins, a minor pool reaches the cytosol by translocating from endosomes, whereas the rest is routed to lysosomes for degradation. Here we demonstrate that part of endocytosed AIP56 is recycled back and released extracellularly through a mechanism requiring phosphoinositide 3-kinase (PI3K) activity but independent of endosome acidification. So far, we have been unable to detect biological activity of recycled AIP56, thereby bringing into question its biological relevance as well as the importance of the recycling pathway.

INTRODUCTION

AIP56 (apoptosis-inducing protein of 56 kDa) is a plasmid-encoded toxin of Photobacterium damselae subsp. piscicida (1), a Gram-negative bacterium that infects economically important fish species (2, 3) and is considered one of the most relevant pathogens in mariculture (2–5). In acute infections, a rapid septicemia develops, causing very high mortalities (2, 4, 6). Histopathological analysis of the diseased animals revealed cytotoxic alterations (4, 7–10) that were found to result from pathogen-induced apoptotic death of macrophages and neutrophils (11) and later associated with the activity of AIP56 (1). Indeed, it has been shown that in infected fish, the toxin is systemically distributed and depletes macrophages and neutrophils by postapoptotic secondary necrosis (12), leading to the impairment of the phagocytic defense of the host and consequently favoring P. damselae subsp. piscicida survival and dissemination. Furthermore, the occurrence of a secondary necrotic process in which the phagocytes undergoing apoptosis lyse and release their cytotoxic contents contributes to the genesis of the infection-associated necrotic lesions (12, 13). These observations, together with the facts that AIP56 is secreted only by virulent P. damselae subsp. piscicida strains and that neutralizing antibodies to the toxin protect fish from P. damselae subsp. piscicida infections (1), established AIP56 as a key virulence factor of P. damselae subsp. piscicida.

AIP56 is the founding member of a continuously growing family of bacterial proteins. Indeed, since we first described AIP56 (1), several open reading frames coding for AIP56 full-length homologues were identified in different organisms, mainly in marine Vibrio species but also in Arsenophonus nasoniae. Interestingly, AIP56 seems to have originated from a fusion of two components: one related to NleC, which is a type III secreted effector present in several enteric pathogenic bacteria (14) associated with human illness and death worldwide (15), and another related to a protein of unknown function from the bacteriophage APSE2. In line with this, we recently found that AIP56 is an AB-type toxin, possessing a catalytic A domain at its N terminus, homologous to NleC, and a B domain involved in binding/internalization into target cells at its C-terminal region, homologous to APSE2 (16). The catalytic domain of AIP56 is a zinc-dependent metalloprotease that, similarly to NleC, cleaves NF-κB, an evolutionarily conserved transcription factor that regulates the expression of inflammatory and anti-apoptotic genes, playing a key role in host responses to microbial pathogen invasion (17). We have shown that during intoxication of sea bass peritoneal leukocytes (sbPL), the catalytic activity of AIP56 results in the depletion of NF-κB p65 (16), and this likely explains the disseminated apoptosis observed during P. damselae subsp. piscicida infections.

Although it is well established that AIP56 is a key virulence factor of P. damselae subsp. piscicida and despite the available knowledge on its structural organization, no information is available concerning the trafficking of AIP56 in intoxicated cells. Since there are several (so far uncharacterized) AIP56 full-length homologues in different bacterial species that are likely to also have crucial roles in virulence, expanding the knowledge on AIP56's intoxication mechanism not only will add to the understanding of P. damselae subsp. piscicida pathogenesis but also may shed light on the pathogenesis of other bacteria producing AIP56-like toxins. The aim of the present study was to define the mechanisms involved in the entry of AIP56 into target cells and to identify its intracellular pathways.

AB toxins are known to target and catalytically modify cytosolic substrates and, to reach the cytosol, have distinct mechanisms of internalization, all involving receptor-mediated endocytosis followed by vesicular trafficking to the site of membrane translocation (18). Short-trip AB toxins, such as diphtheria toxin, CNF1, botulinum neurotoxin, and tetanus neurotoxin, translocate from endosomes following low-pH-induced conformational changes in the toxin molecule (19–21), while others, including Shiga and cholera toxins, are transported retrogradely to the endoplasmic reticulum, from where they translocate the catalytic moiety into the cytosol (22). Prompted by the reduced number of tools (e.g., antibodies) presently available to study AIP56 trafficking in sbPL and also aiming at exploring potential biomedical applications for AIP56, we tested several mammalian cells in intoxication assays. Although P. damselae subsp. piscicida does not grow at 37°C and has a strict salt requirement (4), rendering it unable to infect mammals, here we present evidence that mouse bone marrow-derived macrophages (mBMDM) are susceptible to intoxication by AIP56. Using both sbPL and mBMDM, we have characterized many of the steps involved in the trafficking of the toxin. Our results show that after clathrin-dependent endocytosis, AIP56 translocates from endosomes into the cytosol using a pH-dependent mechanism. Additionally, our data suggest that a pool of toxin follows the endocytic recycling pathway and is released back into the extracellular medium.

MATERIALS AND METHODS

Ethics statement.

This study was carried out in accordance with European and Portuguese legislation for the use of animals for scientific purposes (Directive 2010/63/EU; Decreto-Lei 113/2013). The work was approved by Direcção-Geral de Alimentação e Veterinária (DGAV), the Portuguese authority for animal protection (references 004933 and 2011-02-22).

Reagents.

Concanamycin A (C9705), bafilomycin A1 (B1793), dynasore (D7693), cytochalasin D (C8273), nocodazole (M1404), chlorpromazine (C8138), staurosporine (S4400), phenylmethylsulfonyl fluoride (PMSF) (P7626), pronase E from Streptomyces griseus (P5147), brefeldin A (B5936), cholera toxin (C8052), and lipopolysaccharide (LPS; L4391) were purchased from Sigma-Aldrich. LY294002 (no. 9901) was from Cell Signaling. Hanks balanced salt solution (HBSS), HEPES, and sodium pyruvate were purchased from Invitrogen. Dulbecco's modified Eagle medium (DMEM), Leibovitz's L-15 medium, fetal bovine serum (FBS), and penicillin-streptomycin (P/S) (all Gibco products) and 2-(p-toluidinyl)naphthalene-6-sulfonic acid (TNS; T53) were purchased from Life Technologies. The proteasome inhibitor MG132 (CAS 133407-82-6) was obtained from Calbiochem.

Antibodies and fluorescent probes.

The anti-NF-κB p65 C-terminal domain (c-20) rabbit polyclonal antibody (sc-372), the anti-GAPDH (6C5) (sc-32233), and the anti-c-Myc (9E10) (sc-40) mouse monoclonal antibodies were from Santa Cruz Biotechnology. The anti-V5 (R960-25) and the anti-α-tubulin (32-2500) mouse monoclonal antibodies were from Invitrogen, the anti-early endosomal antigen 1 (EEA1) rabbit polyclonal antibody (E3906) was from Sigma-Aldrich, the anti-lysosomal membrane glycoprotein LAMP-1 mouse monoclonal (ab13523) and the anti-GRP78 Bip rabbit polyclonal (ab32618) antibodies were from Abcam, and the anti-GM130 mouse monoclonal antibody was from BD Biosciences (610822). The anti-Akt rabbit polyclonal antibody (no. 9272) and the anti-phospho-Akt (Ser473) mouse monoclonal antibody (no. 4051) were from Cell Signaling. The anti-sea bass NF-κB p65 rabbit serum was produced using the peptide SIFNSGNPARFVS, located at the C-terminal region of sea bass p65, as the antigen, as described previously (16). Sheep anti-rabbit horseradish peroxidase (HRP)-conjugated secondary antibody (AP311) and sheep anti-mouse HRP-conjugated secondary antibody (AP271) were from The Binding Site; goat anti-rabbit alkaline phosphatase-conjugated secondary antibody (A9919) and goat anti-mouse alkaline phosphatase-conjugated secondary antibody (A2429) were from Sigma-Aldrich. Alexa Fluor 647-transferrin (T23366) was from Invitrogen, Alexa Fluor 594 phalloidin, Alexa Fluor 568 goat anti-rabbit IgG (heavy plus light chain [H+L]), Alexa Fluor 568 goat anti-mouse IgG (H+L), and Alexa Fluor 594 donkey anti-rabbit IgG (H+L) were from Molecular Probes.

Production and fluorescent labeling of recombinant proteins.

In this study, the following recombinant proteins were used: (i) full-length His-tagged AIP56 (recombinant AIP56) (1); (ii) AIP56 carrying a C-terminal Myc or V5 tag followed by a His tag (AIP56Myc and AIP56V5, respectively); (iii) a His-tagged AIP56 metalloprotease mutant (AIP56^AAIVAA) corresponding to a full-length version of the toxin in which key residues for zinc ion coordination and water molecule activation (His¹⁶⁵, Glu¹⁶⁶, His¹⁶⁹, and His¹⁷⁰) were replaced by alanines (16); (iv) a His-tagged AIP56^1–285 with cysteine 262 mutated to serine (AIP56^1–285C262S); (v) a His-tagged AIP56^286–497 with cysteine 298 mutated to serine (AIP56^{286–497C298S}); and (vi) an LF·AIP56 chimeric protein consisting of the amino terminus of anthrax lethal factor (LF^11–263) fused to the AIP56 N-terminal domain (LF^11–263 · AIP56^1–261). Protective antigen (PA), the receptor binding to LF (23), was kindly provided by Cesare Montecucco (University of Padua, Italy). Recombinant AIP56, AIP56^AAIVAA, AIP56^1–285C262S, AIP56^{286–497C298S}, and LF^11–263 · AIP56^1–261 were produced and purified as previously described (16). Briefly, AIP56, AIP56^{286–497C298S}, and LF^11–263 · AIP56^1–261 were purified from the soluble fraction of induced E. coli cells by metal affinity chromatography. This was followed by an anion-exchange chromatography in the case of AIP56 or size exclusion chromatography in the case of AIP56^{286–497C298S}. AIP56^AAIVAA and AIP56^1–285C262S were purified from inclusion bodies by metal affinity chromatography under denaturing conditions, refolded, and subjected to size exclusion chromatography. To obtain AIP56Myc and AIP56V5, pET28aAIP56H+ plasmid (1) was amplified using the forward primer AIP56NdeFw1 (5′-CGCCATATGGCATAACCTTCAATGATGGT-3′) together with AIP56MycXhoRv1 (5′-CGCCTCGAGAAGATCTTCTTCAGAAATAAGTTTCTGTTCATTAATGAATTGTGGCGCGTGGGGAT-3′) or AIP56V5XhoRv1 (5′-CGCCTCGAGCGTAGAATCGAGACCGAGGAGAGGGTTAGGGATAGGCTTACCATTAATGAATTGTGGCGCGTGGGGAT-3′). The PCR products were subcloned into pGEM-T Easy vector (Promega) and cloned into the pET28a(+) plasmid (Promega), yielding the plasmids pET28aAIP56MycH+ and pET28aAIP56V5H+. AIP56Myc and AIP56V5 were expressed in Escherichia coli BL21(DE3) and purified following the protocol used for AIP56 (16). Recombinant proteins were analyzed by SDS-PAGE (see Fig. S1A in the supplemental material), and purities were determined by densitometry of Coomassie blue-stained gels. Protein batches used in this work had a purity of ≥90% (see Fig. S1B in the supplemental material).

Recombinant AIP56 labeled with Alexa Fluor 488 (Alexa 488-AIP56) was prepared using the Alexa Fluor protein labeling kit (A-10235) from Molecular Probes following the manufacturer's instructions.

Determination of protein concentration.

Concentrations of recombinant protein were determined by measuring absorbance at 280 nm and using the extinction coefficient calculated by the ProtParam tool (http://www.expasy.org/tools/protparam.html), using the Edelhoch method (24), but with the extinction coefficients for Trp and Tyr determined by Pace et al. (25). In caspase 3 assays, to normalize the results, the protein concentration in each sample was determined using the Bio-Rad protein assay following the recommended procedure for microtiter plates.

Experimental animals.

The C57BL/6 mice were purchased from Charles River (Madrid, Spain) and bred and housed at the animal facility of the Instituto de Biologia Molecular e Celular (IBMC). The mice were fed sterilized food and water ad libitum. Mice were euthanized by isoflurane anesthesia followed by cervical dislocation. Sea bass (Dicentrarchus labrax) were kept in a recirculating, ozone-treated salt water (25 to 30‰) system at 20 ± 1°C and fed at a ratio of 2% body weight/day. Fish were euthanized with 2-phenoxyethanol (Panreac; >5 ml/10 liters).

Cells.

Mouse bone marrow derived macrophages (mBMDM) were derived from bone marrow of femurs from 4- to 12-week-old C57BL/6 male mice, as previously described (26), and used at day 10 when fully differentiated. Sea bass peritoneal leukocytes (sbPL) were obtained from peritoneal cavities of sea bass, as previously described (27), and used at a density of 2 × 10⁶ cells/ml.

Intoxication assays.

sbPL suspensions or mBMDM monolayers were intoxicated by continuous incubation with recombinant AIP56 (178 nM at 22°C for sbPL and 89 nM at 37°C for mBMDM). Alternatively, cells were pulsed with the toxin for 30 min on ice, washed, and incubated at 22°C (sbPL) or 37°C (mBMDM). In the initial experiments characterizing the susceptibility of mBMDM to AIP56, intoxication was accessed by two different readouts: detection of cytotoxicity through morphological analysis by phase-contrast microscopy and detection of NF-κB p65 cleavage, as previously described (16). In the subsequent experiments characterizing the intracellular trafficking of AIP56, cleavage of NF-κB p65 was used as an indication of the arrival of AIP56 into the cytosol. NF-κB p65 cleavage was assessed by Western blotting after 2 h of incubation with the toxin. The metalloprotease mutant AIP56^AAIVAA was included in some experiments to assess the involvement of the metalloprotease activity in intoxication of mBMDM.

Caspase 3 activity and TUNEL.

The activity of caspase 3 in total cell lysates was determined using a commercial fluorimetric assay (EnzChek caspase 3 assay kit; Molecular Probes). Lysates from cells undergoing apoptosis in response to treatment for 3 h with a 5 μM concentration of the apoptosis inducer staurosporine were used as a positive control. The internucleosomal DNA fragmentation was detected by TUNEL (terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling) labeling of the DNA strand breaks with an in situ cell death detection kit (Roche Diagnostics).

Competition assays.

The competition assays were performed as previously described (16). Briefly, cells were preincubated for 15 min on ice with 7 μM AIP56 A domain (AIP56^1–285C262S) or with 7 μM B domain (AIP56^{286–497C298S}) followed by incubation for 15 min on ice with an 800-fold-lower concentration of recombinant AIP56 (8.75 nM AIP56) in the presence of the competitors. Cells were washed with ice-cold supplemented DMEM and incubated at 37°C for 2 h. NF-κB p65 cleavage was assessed by Western blotting.

Inhibition of trafficking and toxicity.

Cells were pretreated with concanamycin A (10 nM), bafilomycin A1 (10 nM), dynasore (80 μM), chlorpromazine (30 μM), cytochalasin D (4 μM), or LY294002 (10 μM) for 1 h or with nocodazole (33 μM) for 3 h. Results of experiments performed to (i) confirm that the inhibitors did not affect the proteolytic activity of recombinant AIP56 in vitro, (ii) confirm that nocodazole impaired microtubule function and cytochalasin D effectively disrupted actin filaments, (iii) confirm that LY294002 inhibits phosphoinositide 3-kinase (PI3K) activity, and (iv) control the specificities of concanamycin A, dynasore, and chlorpromazine are presented as supplemental material. sbPL were pretreated with the inhibitors at 22°C and mBMDM at 37°C. Cells treated with dynasore were incubated in culture medium lacking FBS. Recombinant AIP56 was added at final concentration of 178 nM (sbPL) or 89 nM (mBMDM) and incubated with the cells for 30 min on ice. Cells were washed and transferred to 22°C (sbPL) or 37°C (mBMDM) and incubated up to 2 h while inhibitory conditions were maintained. Mock-treated cells, cells treated only with toxin, and cells treated only with inhibitors were used as controls. NF-κB p65 cleavage was analyzed by Western blotting.

Detection of endocytosed AIP56.

Cells were pulsed with AIP56V5 or AIP56Myc for 30 min on ice (171 nM for sbPL and 85.5 nM for mBMDM), washed three times with appropriate culture medium, and incubated at 22°C (sbPL) or 37°C (mBMDM). Mock-treated cells were used as controls. When specified, cells were pretreated for 1 h and incubated in the presence of 20 μM MG132 or pretreated and incubated with concanamycin A, bafilomycin A1, dynasore, chlorpromazine, cytochalasin D, or nocodazole as described above. At different time points, cells were washed twice with ice-cold phosphate-buffered saline (PBS) and incubated with 500 μg/ml pronase E in PBS for 10 min on ice to digest the extracellular toxin. Cells treated with the toxin and washed and incubated in PBS were used as controls. The protease was inactivated with PMSF at 1.45 mM. Cells were analyzed by Western blotting for detection of AIP56V5 or AIP56Myc.

Lag time experiments.

Cells were incubated with 178 nM (sbPL) or 89 nM (mBMDM) recombinant AIP56 for 30 min on ice. Medium was replaced by supplemented medium without toxin, and cells were incubated for 2 h at 37°C (mBMDM) or 22°C (sbPL). At various time points, 10 nM concanamycin A was added. The cleavage of NF-κB p65, evaluated by Western blotting, was used as an indication of the arrival of the toxin into the cytosol.

Fluorescence microscopy.

sbPL suspensions and mBMDM cultured on 12-mm glass coverslips were incubated with 178 nM Alexa 488-AIP56 for 30 min on ice. The supernatant was removed and replaced with supplemented medium without toxin, and cells were switched to 22°C (sbPL) or 37°C (mBMDM) for a maximum of 60 min. At different time points, cells were washed once with PBS and processed for immunofluorescence. Cytospin preparations were obtained from sbPL suspensions, fixed with ice-cold 4% (wt/vol) paraformaldehyde in PBS (4% PFA) for 10 min at room temperature (RT) and permeabilized with 0.1% Triton X-100 in PBS for 10 min at RT. mBMDM were either fixed with ice-cold 4% PFA and permeabilized with 0.1% Triton X-100 in PBS (for detection of EEA1, LAMP-1, GM130, and α-tubulin) or fixed with ice-cold methanol for 10 min at RT (for detecting GRP78 Bip). Cells were incubated with blocking buffer (10% FBS in PBS with 0.1% Tween 20 [PBS-T]) for 30 min at RT followed by incubation with the primary antibodies in blocking buffer overnight at 4°C in a humidified chamber. The secondary antibodies diluted in blocking buffer were incubated with cells for 1 h at RT. In experiments with transferrin, mBMDM were incubated in supplemented DMEM without FBS with 178 nM Alexa 488-AIP56 for 20 min on ice, the supernatant was removed and replaced with ice-cold supplemented DMEM containing 6.25 mM Alexa 647-transferrin. Cells were incubated for 10 min on ice, switched to 37°C, and incubated for a maximum of 60 min. At various time points, cells were washed twice with PBS and fixed with ice-cold 4% PFA. Nuclei were counterstained with 0.57 μM DAPI. Samples were subjected to three-dimensional (3D) analysis on a laser scanning confocal microscope (Leica TCS SP5 II running LAS AF 2.6 software; Leica Microsystems, Germany) using a 60×/1.40 numerical aperture (NA) oil immersion objective, with a pixel size of 60 nm and a z step size of 210 nm. Each fluorochrome was imaged by sequential acquisition in order to avoid cross talk between channels. The 3D image stacks were deconvolved with Huygens Professional software (SVI, Netherlands) applying the maximum likelihood estimation restoration algorithm. Restored images were processed and maximum intensity projected with Fiji software (28). The object-based colocalization analysis was performed in Fiji open source software (28) with the Squassh plugin of MosaicSuite developed by MOSAIC Group, MPI-CBG, Dresden, Germany (29, 30). The measurements of colocalizations were determined based on signal using the protocol described in detail by Paul et al. and Rizk et al. (29, 30). Briefly, this involved quantifying colocalization in an intensity-dependent manner by computing the sum of all pixel intensities in one channel in all regions where objects colocalize with objects from the other channel (29, 30).

Recycling assay.

mBMDM were incubated for 30 min on ice in serum-free medium containing 171 nM AIP56V5, transferred to 37°C for 5 min to allow endocytosis of the toxin (pulse), and switched to ice again. Cells were washed three times with ice-cold PBS. The third wash was collected and precipitated with trichloroacetic acid (TCA). After being washed, cells were incubated for 10 min on ice with 125 μg/ml pronase E in PBS to remove extracellular toxin, washed twice with PBS, treated with 1.45 mM PMSF, and washed three times with ice-cold PBS, followed by washing with 250 μl/well serum-free medium containing 0.4 μM AIP56^{286–497C298S} (to prevent re-entry of recycled full-length toxin) and incubation at 37°C in the same medium. The medium from the last wash was collected and precipitated with TCA. At different time intervals, supernatant was collected and precipitated with TCA, and the corresponding cell monolayer was washed twice with PBS and lysed with SDS-PAGE sample buffer. TCA precipitates were resuspended in SDS-PAGE sample buffer, and the equivalent of 12 wells for supernatant and washes and 4 wells for monolayer (pellet) was analyzed by Western blotting for detection of AIP56V5. Membranes were reprobed with an anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody (used as a cytosolic marker). In some experiments, cells were incubated with 10 μM LY294002 or 10 nM concanamycin A for 1 h at 37°C before a pulse with AIP56V5. In this case, the inhibitor was maintained during the entire chase.

pH-induced translocation across the cell membrane.

To address the question of whether an acidic pulse can drive translocation of cell-bound AIP56 across the cell membrane, we followed the protocol first described for diphtheria toxin (31, 32). Different pH values were obtained by adding H₃PO₄ to a buffer containing 0.5 mM MgCl₂, 0.9 mM CaCl₂, 2.7 mM KCl, 1.5 mM KH₂PO₄, 3.2 mM Na₂HPO₄, and 137 mM NaCl. mBMDM were incubated for 30 min on ice in supplemented DMEM with 89 nM recombinant AIP56 in the absence or presence of 10 nM concanamycin A (that was maintained during the entire assay to inhibit normal toxin uptake). Supernatant was removed, and cells were incubated at the different pH for 1 h at 37°C followed by incubation in culture medium (pH 7.4) for further 2 h at 37°C. In all experiments, mock-treated cells and cells treated only with toxin were used as controls. NF-κB p65 cleavage was analyzed by Western blotting.

Black lipid bilayer experiments.

The methods used for black lipid bilayer experiments were described in detail previously (33). The instrumentation consisted of a Teflon chamber with two aqueous compartments connected by a small circular hole, with a surface area of about 0.4 mm². Membranes were formed across the hole by painting on a 1% (wt/vol) solution of diphytanoyl phosphatidylcholine (Avanti Polar Lipids, Alabaster, AL) in n-decane. The aqueous salt solution (150 mM KCl) was buffered with 2 mM CaCl₂ and 10 mM HEPES (pH 7.4 or pH 7.0), MES (morpholineethanesulfonic acid; pH 6.5 or pH 6.0), or CH₃COOK (pH 5.5, pH 5.0, or pH 4.5) added to the cis compartment of the chamber. Reducing conditions were established by addition of 1 mM dithiothreitol (DTT) after the membrane had turned black. Approximately 178 pmol of protein was used for each measurement; purified protein samples mixed 1:1 with cholesterol suspension in water were added to the cis compartment of the chamber after the membrane had turned black. The membrane current was measured with a pair of Ag/AgCl electrodes with salt bridges switched in series with a voltage source and a highly sensitive current amplifier (Keithley 427). The amplified signal was recorded by a strip chart recorder. All measurements were performed at RT, and each experiment was repeated at least twice.

Tryptophan and TNS fluorescence.

Tryptophan and TNS fluorescence at different pH values was analyzed in a PerkinElmer LS45 luminescence spectrometer with excitation of 270 nm and an emission scan of 310 nm to 450 nm and with excitation of 366 nm and an emission scan of 380 to 500 nm, respectively. The following buffers were used: for pH 4.0, 4.5, 5.0, and 5.5, 150 mM NaCl, 100 mM ammonium acetate; for pH 6.0 and 6.5, 150 mM NaCl, 100 mM morpholineethanesulfonic acid (MOPS); for pH 7.0 and 7.5, 150 mM NaCl, 100 mM HEPES (34). TNS was added to each buffer at a final concentration of 150 μM. AIP56Myc was added to a final concentration of 1.5 μM in a final volume of 300 μl. Fluorescence measurements were done after 15 min incubation at RT. For the pH shift experiments, the toxin was incubated with TNS at pH 4.0 for 15 min at RT, and the fluorescence was recorded. The pH was then adjusted to 7.0 by gradual addition of 1 N NaOH, and the fluorescence was recorded again. Since the pH 7.0 was not within the buffering range of ammonium acetate, the pH of the sample was checked following the analysis to confirm that the desired pH value was maintained.

SDS-PAGE and Western blotting.

SDS-PAGE was performed using the Laemmli discontinuous buffer system (35). Prior to loading, the samples were boiled for 5 min in SDS-PAGE sample buffer (50 mM Tris-HCl [pH 8.8], 2% SDS, 0.05% bromophenol blue, 10% glycerol, 2 mM EDTA, and 100 mM DTT). For Western blotting, the proteins were transferred onto nitrocellulose membranes. The efficiency of transfer and the protein loading on the membranes were controlled by staining with Ponceau S. The membranes were blocked for 1 h at RT with 5% skim milk in Tris-buffered saline (TBS) containing 0.1% Tween 20 (T-TBS) followed by incubation for 1 h at RT with the primary antibodies diluted in blocking buffer. Immunoreactive bands were detected using horseradish peroxidase-linked secondary antibodies and enhanced chemiluminescence (ECL) West Dura substrate (Pierce Biotechnology) or alkaline phosphatase-conjugated secondary antibodies and nitroblue tetrazolium–5-bromo-4-chloro-3-indolylphosphate (NBT/BCIP) (Promega). Blots shown present representative results from at least 3 independent experiments. Blots were quantified by densitometry analysis using Fiji software. Loading correction was achieved by dividing the density of the AIP56V5, p65, or cl-p65 band by the respective density of Ponceau S. In the case of the recycling assays, the densities of the GAPDH bands were used instead of Ponceau S. Each graph combines the results of at least three independent experiments.

Statistical analysis.

Statistical analysis of results presented in Fig. 1, 2, 3, 5B, 5C, 6, and 8 and in Fig. S6 and S10 in the supplemental material involved performing one-way or two-way analysis of variance (ANOVA). Normality of the data was tested by the Kolmogorov-Smirnov test; and homogeneity of variances was assessed using the Levene test; P values for individual comparisons were calculated using Tukey's honestly significant difference (HSD) multiple-comparison test. Figures 4 and 5A and Fig. S3, S4, and S7 in the supplemental material present several regression models where the dependent variable is modeled as a function of time. In order to obtain these curves, several models were studied (linear, quadratic, inverse, and logarithmic), and the selected models were based on the quality of the adjustment and, in particular, the residual distribution as well as the coefficient of determination (R²). The analysis of the residuals included the verification of normality as well as the study of extreme values. Analyses were performed using the SPSS software, and significance was set at a P value of <0.05.

FIG 1.

Intoxication of mBMDM with AIP56 leads to NF-κB p65 depletion and apoptosis. (A) Incubation of mBMDM with AIP56 results in apoptotic morphological alterations and cell loss. Phase-contrast microscopy images of mBMDM incubated with 89 nM recombinant AIP56 or AIP56^AAIVAA for the indicated time points are shown. The images were derived from one of three experiments. (B) Incubation of mBMDM with AIP56 results in the appearance of TUNEL-positive and condensed nuclei. Cells were mock treated or incubated with 89 nM recombinant AIP56 or catalytic inactive AIP56^AAIVAA for the indicated time points and processed for the detection of DNA fragmentation by TUNEL (green). Nuclei were counterstained with propidium iodide (red). Images shown are derived from one of three experiments. (C) Incubation of mBMDM with AIP56 results in caspase 3 activation. Cells were incubated with 89 nM recombinant AIP56 or AIP56^AAIVAA for the indicated times, and caspase 3 activity was determined by fluorimetry (with duplicates) using the substrate N-benzoyl-Asp-Glu-Val-Asp-amino-4-methylcoumarin (Z-DEVD-AMC). The results, in relative fluorescence units per nanogram of protein, were converted to fold increase by comparison to values obtained with extracts from mock-treated cells (seven independent experiments; the middle line corresponds to the median, and the tops and bottoms of the boxes represent the first and third quartiles; the circle and asterisk signal extreme observations). The fold increase in caspase 3 activity following treatment with AIP56^AAIVAA (white bars) is not statistically different from 1, whereas that following treatment with recombinant AIP56 (dark gray bars) is significantly different from 1 at all time points (P values for 3, 6, and 9 h: 0.018, 0.008, 0.033, respectively). As expected, a strong activation of caspase 3 was observed in staurosporine-treated cells (positive control, light gray bar; five independent experiments, P = 0.033). Statistical analysis involved performing one sample t test for the hypothesis that the fold increase is equal to 1. (D) Incubation of mBMDM with AIP56 results in NF-κB p65 depletion. Cells were left untreated or incubated with the indicated doses of recombinant AIP56 for 30 min on ice, washed, and chased at 37°C. Cleavage of NF-κB p65 was analyzed by Western blotting (chromogenic detection). The graph shows quantification of the blots. Loading correction was achieved by dividing the density of p65 band by the respective density of Ponceau S. The results were divided by the same constant in order to set to 1 the mean for the control (cells not exposed to recombinant AIP56). Plotted values correspond to the estimated means from three independent experiments. Dose and time dependence of intact p65 levels was tested by two-way ANOVA [dose main effect, P < 0.001; time main effect, P < 0.001; and interaction dose with time, P < 0.001; there was a significant difference between the “no toxin (dose 0)” group and the other groups: P < 0.001 (Tukey′s HSD)].

FIG 2.

AIP56 endocytosis and toxicity require clathrin and dynamin. (A) AIP56 endocytosis is prevented by Cpz and Dyn. mBMDM were pretreated with Cpz or Dyn, pulsed with AIP56V5 in the presence of the inhibitors for 30 min on ice, washed, and incubated at 37°C in culture medium with inhibitors. After 2 min, surface-exposed toxin was removed with PronE, and the intracellular pool was detected by Western blotting (chromogenic detection). (B) AIP56-mediated cleavage of NF-κB p65 is inhibited by Cpz and Dyn. sbPL or mBMDM were pretreated with Cpz or Dyn, pulsed with recombinant AIP56 for 30 min on ice in the presence of the inhibitor, shifted to 22°C (sbPL) or 37°C (mBMDM), and incubated for 2 h. Cleavage of p65 was assessed by Western blotting (chromogenic detection). (C) Intoxication of mBMDM by AIP56 is inhibited by the toxin's B domain AIP56^{286–497C298S}. Cells were incubated with 7 μM AIP56^1–285C262S (AIP56Nterm) or AIP56^{286–497C298S} (AIP56Cterm) for 15 min on ice, followed by further 15 min incubation with an 800-fold-lower concentration of recombinant AIP56 in the presence of the competitor. Cells were washed and incubated at 37°C for 2 h. Cleavage of NF-κB p65 was assessed by Western blotting (chromogenic detection). The graphs show the quantification of the blots. Loading correction was achieved by dividing the density of AIP56V5 (A) or p65 (B and C) by the respective density of Ponceau S. The results used for each graph were divided by the same constant in order to set to 1 the mean of the control (cells exposed to toxin only [A] or mock-treated cells [B and C]). Each graph combines the results of three independent experiments (means ± standard deviations [SD]). The significance of differences was tested by one-way ANOVA (all P values were <0.001). P values for individual comparisons were calculated using Tukey's HSD multiple-comparison test and are indicated in the graphs.

FIG 3.

AIP56 intoxication is independent of actin and microtubules. (A) AIP56 endocytosis is not inhibited by CytoD or Noco. mBMDM were treated with CytoD or Noco before incubation with AIP56V5 in the presence of the inhibitors for 30 min on ice (pulse), washed, and incubated for 2 min at 37°C in culture medium with inhibitors. Extracellular toxin was removed with PronE, and the intracellular pool was detected by Western blotting (anti-V5 antibody, chromogenic detection). (B) The arrival of AIP56 into the cytosol is not prevented by CytoD or Noco. sbPL or mBMDM were treated with CytoD or Noco at the appropriate temperature prior to incubation with recombinant AIP56 for 30 min on ice in the presence of the inhibitor. Cells were shifted to 22°C (sbPL) or 37°C (mBMDM) and incubated for 2 h while the inhibitory conditions were maintained. Cleavage of NF-κB p65 was assessed by Western blotting (chromogenic detection). The graphs show the quantification of the blots. Loading correction was achieved by dividing the density of AIP56V5 (A) or p65 (B) by the respective density of Ponceau S. The results used for each graph were divided by the same constant in order to set to 1 the mean of the control (cells exposed to toxin only [A] or mock-treated cells [B]). Each graph combines the results of three independent experiments (mean ± SD). The significance of differences was tested by one-way ANOVA (P < 0.001 [A], P < 0.012 [B, mBMDM], and P < 0.001 [B, sbPL]). P values for individual comparisons were calculated using Tukey's HSD multiple-comparison test and are indicated in the graphs.

FIG 5.

After endocytosis, AIP56 is recycled back to the extracellular medium by a mechanism requiring PI3K activity but independent of endosomal acidification. (A) mBMDM were left untreated or incubated with AIP56V5 on ice for 30 min, followed by 5 min incubation at 37°C (pulse). Cells were washed, extracellular toxin was removed with PronE, and cells were washed again and chased at 37°C for different time intervals. Cell-associated (pellet) and recycled (supernatant) AIP56V5 was detected by Western blotting (anti-V5 antibody; chromogenic detection). To control the efficacy of PronE treatment, we analyzed the presence of the toxin in the last wash before PronE treatment (wash BP) and in the last wash after PronE treatment (wash AP). Samples (washes, supernatants, and pellets) from 12 wells were pooled and the equivalents of 12 wells for washes/supernatants or 4 wells for pellets were loaded in the gels. GAPDH was used as a cytosolic marker. The graphs show the quantification of the recycled or cell-associated AIP56V5 relative to the total internalized protein (lane 3) and combine results from three independent experiments (represented by different symbols). The logarithmic regression model represents the percentage of recycled or cell associated AIP56V5 as a function of time (recycled R² = 0.451; cell-associated R² = 0.679). (B) LY inhibits the recycling of AIP56 into the extracellular medium but not its toxicity. mBMDM were left untreated or treated with LY before the recycling or toxicity assays in the presence of LY. (I) The recycling protocol was the same as described for panel A. The graph shows the quantification of the blots and combines results from four independent experiments (the middle line in each box corresponds to the median, and the tops and bottoms of the boxes correspond to the first and third quartiles; whiskers represent the smallest and largest observed values). Values correspond to percentage of recycled AIP56V5 relative to the total internalized protein (lane 3). Statistical analysis involved performing two-way ANOVA (main effect of inhibitor, P < 0.001; main effect of time, P = 0.073; inhibitor and time interaction, P = 0.059). (II) Toxicity (cleavage of NF-κB) was assessed and quantified as described for Fig. 2B. The significance of differences was tested by one-way ANOVA (three independent experiments; mean ± SD; P < 0.001). P values for individual comparisons were calculated using Tukey's HSD multiple-comparison test and are indicated in the graph. (C) Recycling of AIP56 into the extracellular medium is not inhibited by ConcA. mBMDM were treated with ConcA or left untreated before the recycling assay, as described for panel A, but in the presence of ConcA. The graph shows the quantification of the blots and combines results from three independent experiments (the middle line in each box corresponds to the median, and the tops and bottoms of the boxes correspond to the first and third quartiles; whiskers represent the smallest and largest observed values). Values correspond to percentage of recycled AIP56V5 relative to the total internalized protein (lane 3). Statistical analysis involved performing two-way ANOVA (main effect of inhibitor, P = 0.180; main effect of time, P = 0.784; inhibitor and time interaction, P = 0.942).

FIG 6.

AIP56 translocation into the cytosol is dependent on endosome acidification. (A) AIP56 endocytosis is not inhibited by ConcA or BafA1. mBMDM were treated with BafA1 or ConcA before incubation with recombinant AIP56 in the presence of the inhibitors for 30 min on ice, washed and incubated for 2 min at 37°C in culture medium with inhibitor. Extracellular toxin was removed with PronE and the intracellular pool was detected by Western blotting (anti-V5 antibody; chromogenic detection). (B) ConcA and BafA1 prevent the AIP56-dependent cleavage of p65. sbPL or mBMDM were treated with the indicated inhibitor before incubation with recombinant AIP56 in the presence of the inhibitor for 30 min on ice. Cells were washed, shifted to 22°C (sbPL) or 37°C (mBMDM) and incubated for 2 h while maintaining the inhibitory conditions. Cleavage of p65 was assessed by Western blotting (chromogenic detection). (C) Time course of AIP56 translocation. sbPL or mBMDM were left untreated or incubated with recombinant AIP56 for 30 min on ice, washed and transferred to 22°C (sbPL) or 37°C (mBMDM). ConcA was added at the indicated time points, and after 2 h, cleavage of p65 was assessed by Western blotting (chromogenic detection). The graphs show quantification of the blots. Loading correction was achieved by dividing the density of AIP56V5 (A) or p65 (B and C) by the respective density of Ponceau S. The results used for each graph were divided by the same constant in order to set to 1 the mean of the control (cells exposed to toxin only [A], mock-treated cells [B], or ConcA at 0 min [C]). Each graph combines the results of three independent experiments (mean ± SD). The significance of differences was tested by one-way ANOVA (P < 0.001 [A], P < 0.01 [B, mBMDM] P < 0.001 [B, sbLP], P < 0.001 [C, mBMDM], and P = 0.126 [C, sbPL]). P values for individual comparisons were calculated using Tukey's HSD multiple comparisons and are indicated in the graphs (in panel C, the P values are for comparisons with cells treated only with toxin).

FIG 8.

An acidic pulse can drive translocation of AIP56 across the cell membrane. mBMDM treated with ConcA (that was maintained during the entire assay to inhibit normal toxin uptake) or left untreated were incubated on ice with recombinant AIP56. After 30 min, medium was removed, and cells were incubated for 1 h at 37°C with buffers at the indicated pH values followed by incubation in culture medium at pH 7.4 for 2 h. Cleavage of p65 was assessed by Western blotting (chromogenic detection). The graph shows the quantification of the blots (mean ± SD; three independent experiments). Loading correction was achieved by dividing the density of p65 band by the respective density of Ponceau S. The results were divided by the same constant in order to set to 1 the mean of the control (mock-treated cells, pH 7.0). For each pH, significance of differences was assessed by one-way ANOVA (P < 0.01). P values for the comparisons with cells treated only with toxin were calculated using Tukey's HSD multiple-comparison test and are indicated in the graph.

FIG 4.

After endocytosis, AIP56 localizes in early endosomes and then is routed to the recycling compartment. (A) Representative confocal images of mBMDM pulsed with Alexa 488-AIP56 (green) for 30 min on ice, washed, and chased at 37°C. At the indicated time points, cells were fixed and processed for immunofluorescence for detection of EEA1, LAMP-1, GM130, or GRP78 Bip. The entire recycling pathway was labeled with a continuous incubation with Alexa 647-transferrin (TF). Nuclei were counterstained with DAPI (blue). A merge of the different channels is shown (images of individual channels are available in Fig. S8 in the supplemental material), and insets are magnified views of boxed areas. (B) Quantification of Alexa 488-AIP56 colocalization with EEA1 and TF. For each marker, the percentage of colocalization was determined for several cells (n ≥ 15) obtained from at least three independent experiments. Each point in the plot corresponds to the average of an independent experiment. Linear regression models were developed for each of the markers. In the case of TF, the model represents colocalization with TF as a function of the logarithm of time (R² = 0.71), whereas for EEA1, it represents the logarithm of EEA1 as a function of the logarithm of time (R² = 0.91). For the remaining components, the observed colocalization and its variation with time were negligible, so that the best linear model was an almost horizontal line (not shown).

RESULTS

AIP56 induces depletion of NF-κB p65 and apoptosis in mBMDM.

The lack of appropriate and established tools for studying intracellular trafficking in sbPL led us to search for an alternative cell model. Although mammals are not susceptible to infection by P. damselae subsp. piscicida, likely due to temperature and osmolarity restrictions (4), we decided to evaluate the susceptibility of different mammalian cells to AIP56. Thus, we tested macrophage (J774A.1, Raw264.7, and THP-1) and epithelial (HeLa, CHO, and Vero) cell lines (data not shown) as well as mBMDM. Toxicity was observed only in the latter (Fig. 1). When these cells were treated with recombinant AIP56, morphological alterations were evident from 6 to 24 h of incubation (Fig. 1A). These alterations were dose dependent (see Fig. S2 in the supplemental material) and required the metalloprotease activity of the toxin, since they did not occur after incubation with the AIP56 metalloprotease mutant AIP56^AAIVAA (Fig. 1A). The cellular alterations included shrinkage, rounding of the cells, and membrane blebbing, as well as detachment of almost all cells after 24 h of incubation (Fig. 1A), suggesting the occurrence of an apoptotic process. Assessment of nuclear DNA fragmentation by TUNEL and measurement of caspase 3 activity confirmed that intoxicated mBMDM were undergoing apoptosis (Fig. 1B and C). As expected, cells treated with AIP56^AAIVAA revealed no signs of apoptosis (Fig. 1B and C). Western blotting revealed that intoxication of mBMDM by recombinant AIP56 resulted in time- and dose-dependent decreases of NF-κB p65 protein levels (Fig. 1D). The correlation between p65 cleavage and the occurrence of apoptosis has not been addressed, and therefore, the question of whether p65 depletion per se is sufficient to induce death or whether there are other host cell proteins targeted and implicated in cytotoxicity needs to be clarified. The presence of the p65 cleavage fragment (cl-p65) is time dependent, since it was detected in cells up to 90 min after intoxication but not in cells subjected to longer incubation times (Fig. 1D). Similarly, in sbPL, the p65 cleavage fragment was detected up to 4 h but was no longer visible after intoxication for 6 h (see Fig. S3 in the supplemental material). Using the proteosomal inhibitor MG132, we found that degradation of the p65 cleavage fragment originated by AIP56 depends on proteosomal activity, in agreement with what has been reported for the p65 cleavage fragment resulting from the activity of the type III effector NleC (36). It is possible that the faster degradation of the cleavage fragment observed in mBMDM results from a larger activity of the proteasomal machinery in mBMDM than in sbPL.

AIP56 is endocytosed through a clathrin-dependent mechanism involving interaction of the C-terminal region of the toxin with cell surface components.

The observation that AIP56 cleaves NF-κB in sbPL and mBMDM indicates that the toxin is able to reach the cytosol of these cells. Here, we used fluorescence microscopy and biochemical approaches to investigate the entry mechanisms used by the toxin. After 5 min incubation with sbPL or mBMDM, Alexa 488-AIP56 was observed in vesicle-like structures (see Fig. S4A in the supplemental material), suggesting that it had been endocytosed. Western blotting confirmed the internalization of the toxin (see Fig. S4B in the supplemental material). A kinetic analysis of the entry in sbPL revealed a marked decrease of intracellular toxin over time (see Fig. S4C in the supplemental material), which was not prevented by the proteosomal inhibitor MG132 (see Fig. S4C in the supplemental material). To enter host cells, bacterial toxins are known to use clathrin-dependent and/or clathrin-independent endocytic mechanisms. Dynamin is a GTPase involved in endocytic membrane fission that is known to be involved in clathrin-dependent endocytosis, although it may also participate in clathrin-independent endocytosis pathways (37). To investigate if AIP56 endocytosis is dependent on clathrin and/or dynamin, we inhibited their functions using the specific cell-permeative inhibitors chlorpromazine (Cpz) (38) and dynasore (Dyn) (39), respectively. In vitro tests confirmed that Cpz and Dyn did not affect AIP56 metalloprotease activity (see Fig. S5 in the supplemental material), and phagocytosis assays confirmed that they were acting specifically (see Fig. S6A in the supplemental material). To test the effect of Cpz and Dyn in AIP56 endocytosis, we treated mBMDM with the inhibitors or left the cells untreated before pulsing with recombinant toxin for 30 min on ice. Cells were washed and transferred to 37°C, and after a 2-min chase, extracellular toxin was removed with pronase E (PronE). As shown in Fig. 2A, endocytosed AIP56 was detected in control cells but not in cells pretreated with Cpz or Dyn. The inhibitory effect of Cpz and Dyn upon AIP56 endocytosis together with the observation that the AIP56-dependent cleavage of p65 was prevented by pretreatment with Dyn and was strongly inhibited by pretreatment with Cpz (Fig. 2B) suggests that a significant pool of AIP56 is endocytosed through a dynamin- and clathrin-dependent pathway. Nevertheless, the partial inhibition of toxicity observed after pretreatment with Cpz indicates that a fraction of toxin molecules may be internalized through a dynamin-dependent mechanism not requiring clathrin. Previous studies with sbPL showed that the entry of the toxin involves specific interaction of its C-terminal region with cell surface components (16). To investigate if the C-terminal region of AIP56 also plays a role in the entry in mBMDM, we performed a competition assay testing the ability of AIP56^{286–497C298S} (corresponding to the toxin's B domain) (16) to inhibit the entry of the toxin into the cytosol. AIP56^1–285C262S, corresponding to the toxin's A domain (16), was used as a control. Preincubation with AIP56^{286–497C298S} before intoxication with recombinant AIP56 inhibited p65 cleavage, whereas no inhibition was observed when AIP56^1–285C262S was used as a competitor (Fig. 2C). These results indicate that the entry of AIP56 into mBMDM involves interaction of AIP56 C-terminal domain with cell surface receptors.

Microtubules and actin dynamics are dispensable for intoxication by AIP56.

Functional actin dynamics is required for several endocytic processes (40). To determine if internalization of AIP56 depends on actin, we disrupted microfilaments with cytochalasin D (CytoD) (41) (see Fig. S6A and B in the supplemental material) and addressed whether disruption inhibited intoxication. Treatment with CytoD did not inhibit AIP56 endocytosis (Fig. 3A) or AIP56-dependent p65 cleavage (Fig. 3B), indicating that intoxication is independent on actin dynamics. The involvement of microtubules in AIP56 uptake was investigated using the microtubule depolymerizing agent nocodazole (Noco) (42). Although Noco effectively disrupted microtubules (see Fig. S7A and B in the supplemental material), it did not affect AIP56 endocytosis (Fig. 3A) or AIP56-dependent p65 cleavage (Fig. 3B), indicating that AIP56 exploits a microtubule-independent intoxication pathway.

Upon endocytosis, AIP56 localizes in early endosomes and then is routed to the recycling endocytic compartment.

To further characterize the intracellular trafficking of AIP56, cells were pulsed with Alexa 488-AIP56 (green) for 30 min on ice, washed, and transferred to 37°C. At different chase times, cells were processed for immunofluorescence to assess the colocalization of toxin with established markers of intracellular compartments. Colocalization experiments are reported only for mBMDM, because most of the commercially available antibodies and probes are not validated or do not work on fish cells. After a 2-min chase, Alexa 488-AIP56 was observed in peripheral endocytic vesicles negative for all the markers tested, but shortly after (after a 5-min chase), part of the toxin-containing vesicles were positive for the early endosome marker EEA-1 (Fig. 4; also, see Fig. S8A in the supplemental material). The colocalization with EEA-1 reached a maximum at 15 to 20 min and afterwards markedly decreased, being minimal at 60 min (Fig. 4; also, see Fig. S8A in the supplemental material). These results indicate that after endocytosis, AIP56 localizes in early endosomes, where it remains for 20 to 30 min. To define the compartment where AIP56 was delivered after leaving early endosomes, we looked for colocalization of the toxin with the late endosome/lysosome marker LAMP-1, the Golgi apparatus marker GM130, or the endoplasmic reticulum marker GRP78 Bip. Alexa 647-transferrin was used as a marker of the recycling route. In macrophages (43–48), similar to what occurs in other cells (49), transferrin is recycled back to the cell surface using two kinetically distinguishable mechanisms of recycling (fast and slow recycling). This supports the use of transferrin as a marker for the recycling pathway in macrophages and in, particular, in mBMDM. Colocalization of Alexa 488-AIP56 with Alexa 647-transferrin was evident at 5 min and was maintained during the remaining chase (Fig. 4; also, see Fig. S8B in the supplemental material), suggesting that after leaving early endosomes, the toxin was delivered to the recycling endocytic compartment. In agreement with this hypothesis, Alexa 488-AIP56 was never observed to colocalize with LAMP-1 (Fig. 4; also, see S8C in the supplemental material) or markers of Golgi apparatus and endoplasmic reticulum (Fig. 4; also, see Fig. S8D and E, respectively, in the supplemental material).

Shortly after endocytosis, a pool of AIP56 is recycled back to the extracellular medium through a mechanism requiring PI3K activity but independent on endosome acidification.

The observation that AIP56 was delivered to the recycling endocytic compartment following endocytosis, suggested that the toxin was being recycled back to the extracellular medium. To investigate this hypothesis, we allowed mBMDM to bind AIP56V5 and internalize it for 5 min (pulse), removed the extracellular toxin with PronE, chased the cells at 37°C and analyzed the presence of the toxin inside cells and in the extracellular medium at different time points. The absence of AIP56V5 in the last wash after PronE treatment (Fig. 5A, lane 5) confirmed that extracellular toxin was removed before chasing. After 2 min chase, AIP56V5 was already detected in the extracellular medium (Fig. 5A, lane 7). Extracellular AIP56V5 increased over time (Fig. 5A, lanes 7 to 10) and was paralleled by a decrease in cell-associated toxin (Fig. 5A, lanes 12 to 15), indicating that during chase, intracellular AIP56 was being recycled to the extracellular medium. Since PI3K, an evolutionarily conserved enzyme complex which phosphorylates phosphatidylinositol lipid substrates into 3-phosphoinositides (50, 51), has been shown to play a key role in regulating the recycling pathway (52–56), we analyzed the involvement of PI3K in AIP56 recycling. For this, we performed recycling assays in the presence of LY294002 (LY), a reversible paninhibitor of PI3K activity (57, 58). The effect of LY upon PI3K activity was controlled by confirming LY-induced inhibition of the phosphorylation of Akt, a downstream target of PI3K (59) (see Fig. S9 in the supplemental material). In the presence of LY, recycling of AIP56 was strongly inhibited (Fig. 5B, compare lanes 6 and 7 with lanes 10 and 11, respectively), and an increase in cell-associated toxin was observed (compare lanes 12 and 13 with lanes 14 and 15, respectively), suggesting that recycling of AIP56 is dependent on PI3K activity. Because PI3K is involved in a wide range of cellular processes (50) including regulation of the endocytic pathway we also analyzed the effect of LY in AIP56 toxicity. As shown in Fig. 5B, no differences in AIP56-dependent p65 cleavage were observed in cells treated with LY compared to those incubated with the toxin in the absence of the inhibitor, indicating that the AIP56 intoxication pathway is independent of PI3K activity. We also investigated if recycling of AIP56 was dependent on endosome acidification, because it is reported that the exposure of ligand-receptor complexes internalized by receptor-mediated endocytosis to low endosomal pH is required for the efficient sorting of several cargos and recycling of receptors (52, 60). The recycling experiment was repeated in the presence of concanamycin A (ConcA), an inhibitor of the endosomal vacuolar ATPase pump responsible for endosome acidification (61). Experiments aiming at controlling the specificity and efficacy of ConcA in inhibiting the endosomal vacuolar ATPase pump are presented in Fig. S10 in the supplemental material. Recycling of AIP56V5 was not inhibited by ConcA (Fig. 5C, compare lanes 6 and 7 with lanes 10 and 11, respectively), suggesting that it does not require low endosomal pH.

Endosome acidification is required for AIP56 intoxication.

The indication that following endocytosis, AIP56 entered early endosomes and that after 20 to 30 min it trafficked into the endocytic recycling compartment suggests that in order to reach its cytosolic target, this toxin translocates from early or recycling endosomes. AB toxins that translocate from endosomes (e.g., diphtheria, tetanus, and botulinum toxins) are inhibited by agents that prevent acidification of endosomes/lysosomes (22, 62, 63), because the trigger for translocation is the low endosomal pH. To clarify the mechanism involved in the translocation of AIP56 into the cytosol, we used bafilomycin A1 (BafA1) and ConcA, two potent inhibitors of the endosomal vacuolar ATPase pump and, thus, of the acidification of early and late endosomes as well as lysosomes (61). These compounds did not affect the in vitro catalytic activity of AIP56 (see Fig. S5 in the supplemental material) or its endocytosis (Fig. 6A). However, when mBMDM or sbPL were pretreated with either inhibitor, AIP56-dependent cleavage of p65 was prevented (Fig. 6B), indicating that arrival of the toxin at the cytosol is dependent on endosome acidification. To gain insight into the time course of AIP56 translocation, we took advantage of the inhibitory effect of ConcA. As shown in Fig. 6C for mBMDM, ConcA markedly inhibited p65 cleavage when added up to 15 min following AIP56 treatment. However, addition of the inhibitor 30 min after toxin treatment resulted in only partial inhibition (not statistically significant), and afterwards, the addition of ConcA had no effect on p65 degradation. These results indicate that at 30 min incubation, some AIP56 is no longer susceptible to the effect of ConcA, suggesting that it has already translocated into the cytosol to exert its toxic effect. A similar effect was observed for sbPL, although in this case, the differences did not reach statistical significance.

At acidic pH, AIP56 undergoes reversible conformational changes and interacts with artificial lipid bilayer membranes.

Considering that toxins that translocate from endosomes undergo conformational rearrangements at endosomal pH (34), pH-induced structural changes in AIP56 were monitored by analyzing TNS fluorescence (TNS is a commercial probe that is nonfluorescent in water and becomes fluorescent when bound to hydrophobic regions of a protein [64]). Recombinant toxin was incubated with TNS at different pH values for 15 min, and fluorescence was analyzed. As shown in Fig. 7A, while TNS fluorescence at pH 6.0 and above was not higher than background levels, at pH 5.5, fluorescence markedly increased. The increase continued at pH 5.0, 4.5, and 4.0 and was found to be reversible, since returning the toxin that had been incubated at pH 4.0 to pH 7.0 resulted in a strong decrease in fluorescence intensity (Fig. 7B). These results indicate that AIP56 exhibits a reversible increase in hydrophobicity when exposed to an acidic (pH 5.5 to 4.0) environment. Because TNS is an “external” probe of protein hydrophobicity, the intrinsic fluorescence of tryptophan residues was also analyzed. Changes in tryptophan fluorescence can be used to detect protein conformational alterations, because those can be quenched in the presence of an aqueous solvent. As shown in Fig. 7C, AIP56 tryptophan fluorescence decreases as pH declines, suggesting that under acidic conditions tryptophan-containing domains move into more aqueous environments or that hydrophobic pockets move away from tryptophans. Following the observation that AIP56 undergoes pH-induced conformational changes, we used black lipid bilayers (33) to determine whether the toxin is able to interact with lipid membranes and whether the observed pH-induced conformational changes play a role in the interaction. When tested at neutral pH values (pH 7.0 to 7.4), recombinant AIP56 failed to exhibit membrane activity in black lipid membranes. However, acidification of the aqueous phase at the side of the cis compartment triggered membrane activity (Fig. 7D). Accordingly, when recombinant AIP56 was added to the cis chamber at acidic pH (6.5 to 4.5), membrane activity was also triggered after a 150-mV pulse (Fig. 7E), leading to an undefined but stepwise increase in membrane conductivity (Fig. 7E, left), which was followed by membrane rupture. A very stable signal was observed when the voltage was kept to 50 mV (Fig. 7E). AIP56^AAIVAA behaved similarly to AIP56 (Fig. 7F), indicating that mutations in the metalloprotease signature did not impair AIP56's membrane activity. The observed current fluctuations were irregular and inhomogeneous, indicating that interaction of AIP56 with bilayer membranes does not lead to the formation of regular channels comparable to the ones formed by other AB toxins, such as anthrax toxins (65, 66) or Clostridium botulinum C2 (65, 66). The current fluctuations induced by AIP56 are similar to the ones reported to occur with the C. botulinum and Clostridium limosum C3 toxins (67) as well as with the Clostridium difficile TcdB and TcdA toxins (68, 69) and may result from formation of transient channels.

FIG 7.

Under acidic conditions, AIP56 undergoes reversible conformational changes and interacts with lipid bilayer membranes. (A) TNS analysis of pH-induced hydrophobic transitions in AIP56Myc. The toxin was incubated with TNS at the indicated pH values, and fluorescence was analyzed. (B) TNS fluorescence analysis following a pH shift. AIP56Myc was incubated with TNS at pH 4.0, and fluorescence was analyzed, as described for panel A. The pH was then adjusted to 7.0 (shifted pH), and the TNS emission spectrum was recorded again. AIP56Myc incubated with TNS at pH 7.0 was used as a control. (C) Intrinsic tryptophan fluorescence of AIP56Myc incubated at pH 7.5 or pH 4.0. In experiments A, B, and C, fluorescence intensities were determined by averaging two readings, and background fluorescence (TNS plus buffer alone for TNS analysis or buffer alone for tryptophan) was subtracted from the fluorescence of each experimental sample. Results are expressed as the averages (black line) ± SD (dashed lines) from four independent experiments. (D) Current recording of a diphytanoyl phosphatidylcholine/n-decane membrane in the presence of recombinant AIP56. The applied membrane potential was 50 mV. Initial experimental conditions consisted of 150 mM KCl, 2 mM CaCl₂, 1 mM DTT, 10 mM HEPES (pH 7.4). Addition of recombinant AIP56 mixed 1:1 with cholesterol suspension in water to the cis compartment of the chamber (left; arrow) had no effect on membrane conductivity. Acidification of the aqueous phase at the cis compartment by addition of 10 mM CH₃COOK (pH 4.6) (right; arrow) triggered membrane activity. The experiment was repeated three times. (E) Membrane activity could also be triggered by a 150-mV pulse. A current recording of a diphytanoyl phosphatidylcholine/n-decane membrane in the presence of recombinant AIP56 mixed 1:1 with cholesterol suspension in water is shown. Measurements were performed in 150 mM KCl, 2 mM CaCl₂, 10 mM MES (pH 6.0), and under these conditions, membrane activity was triggered by a 150-mV pulse (B, left; arrow), which finally resulted in membrane rupture. A very stable signal could be observed when only 50 mV was applied to a diphytanoyl phosphatidylcholine/n-decane membrane in the presence of recombinant AIP56 mixed 1:1 with cholesterol suspension in water (B, right). The experiment was repeated three times. (F) Mutations in the metalloprotease signature do not impair the interaction of AIP56 with lipid bilayer membranes. A current recording of a diphytanoyl phosphatidylcholine/n-decane membrane in the presence of AIP56^AAIVAA mixed 1:1 with cholesterol suspension in water is shown. Measurements were performed in 150 mM KCl, 2 mM CaCl₂, 10 mM MES (pH 6.0). Membrane activity was triggered under these conditions by a 150-mV pulse (left; arrow) and stabilized by lowering the applied voltage to 50 mV (right; arrow). The experiment was repeated twice.

Extracellular acidification triggers translocation of cell surface-bound AIP56 into the cytosol.

The observation that inhibitors which prevent endosome acidification inhibited AIP56 toxicity, that AIP56 undergoes pH-induced conformational changes and is able to interact with black lipid bilayers at low pH prompted us to test if a low-pH pulse could drive the translocation of plasma membrane-bound AIP56 into the cytosol, similar to what has been reported for several other bacterial toxins that translocate from endosomes through a pH-dependent mechanism (31, 32). ConcA was maintained during the entire experiment to block endocytic translocation of AIP56. mBMDM were incubated with recombinant AIP56 for 30 min on ice to allow toxin binding and then subjected to a pulse at pH 5.5 or 5.0 (pH 7.0 as a control) at 37°C. Afterwards, cells were incubated in supplemented culture medium at pH 7.4 for 2 h at 37°C, washed, and analyzed by Western blotting for detection of p65 cleavage. In the absence of an acidic pulse, ConcA inhibited AIP56-dependent cleavage of p65 (Fig. 8). However, the inhibitory effect of ConcA was abolished when mBMDM with surface-bound AIP56 were exposed to pH 5.0 or 5.5 (Fig. 8), indicating that an acidic pulse is sufficient to induce AIP56 translocation across the plasma membrane of targeted cells.

DISCUSSION

In this study, we found that mBMDM are susceptible to AIP56 intoxication and, using both sbPL and mBMDM, defined key steps involved in the entry and intracellular trafficking of the toxin.

Although P. damselae subsp. piscicida is not able to infect mammals, our data show that, similar to what has been recently observed in sbPL (16), intoxication of mBMDM by AIP56 results in NF-κB p65 depletion and apoptosis. It is well known that uncontrolled activation of NF-κB is associated with several human pathologies, including inflammatory diseases and cancers (70–72), and thus, apart from its intrinsic biologic interest, the observation that AIP56 acts on mammalian cells confers considerable biotechnological potential to the toxin.

Similar to diphtheria toxin, CNF1, botulinum neurotoxin, and tetanus neurotoxin (73–75), AIP56 is a single-chain AB toxin that reaches the cytosolic compartment to exert its activity (16). Here, using sbPL and mBMDM, we found that shortly after incubation with cells, AIP56 localizes in intracellular vesicular compartments initially scattered in the cytoplasm that over time concentrate in the perinuclear area, indicating that binding of AIP56 to a still-unidentified cell surface receptor(s) is followed by endocytosis into a vesicular/endosomal compartment, from which it escapes to reach the cytosol. As shown in sbPL (16), the binding of AIP56 to mBMDM occurs by interaction of the C-terminal region of the toxin with the host cell surface, suggesting that the toxin recognizes a conserved receptor. Our results indicate that following receptor binding, AIP56 endocytosis likely occurs through a clathrin-dependent mechanism, since inhibition of the formation of clathrin-coated pits blocked AIP56 internalization. In agreement with this, inhibition of the large GTPase dynamin, known to play an important role in clathrin-mediated endocytosis by participating in the constriction and subsequent budding of coated pits (37, 76–78), also blocked AIP56 entry. Clathrin-dependent endocytosis includes several internalization pathways, all relying on the use of the coat protein clathrin but differing with regard to the requirement for other proteins, including actin (79). Findings of the present study indicate that the actin cytoskeleton is not involved in AIP56 endocytosis, since depolymerization of F actin by CytoD (41) did not affect entry or p65 cleavage by the toxin. This is similar to what has been reported for transferrin, a classical marker of clathrin-mediated endocytosis, which in several cell types has been shown to enter through a pathway independent of actin (80), and also for diphtheria toxin, which is known to enter cells via a clathrin-dependent mechanism that is actin independent (81, 82).

It has traditionally been thought that molecules internalized via clathrin-dependent endocytosis are either recycled back to the plasma membrane or degraded via the lysosomal pathway (79). However, there are toxins, like Shiga and cholera toxins, that upon internalization via clathrin-coated pits traffic retrogradely to the Golgi apparatus (83). Several lines of evidence indicate that upon endocytosis, AIP56 follows the endocytic recycling route and is released into the extracellular medium. (i) Within 5 min after endocytosis, AIP56 localizes in early endosomes (positive for EEA-1 and Alexa 647-transferrin), where it remains for 20 to 30 min, and then is routed to the recycling endocytic compartment (identified by labeling with Alexa 647-transferrin). (ii) Using a recycling assay in which we allowed cells to internalize AIP56, removed extracellular toxin, and then monitored the appearance of recycled toxin in the extracellular medium, we observed that part of the endocytosed AIP56 is recycled back to the extracellular compartment. Recycled toxin is already detected at 2 min chase (following a pulse of 30 min on ice plus 5 min at 37°C), and its amount increases over time, while the amount of cell-associated toxin decreases. The decrease in intracellular toxin is not inhibited by the proteasomal inhibitor MG132, and no colocalization of AIP56 with LAMP-1 was detected in our experiments, suggesting that this decrease results not from proteasomal or lysosomal degradation but rather from recycling of the toxin back into the extracellular medium. (iii) The appearance of recycled AIP56 in the extracellular medium is dependent on PI3K, a kinase that plays a key role in the endocytic recycling of several molecules, including transferrin (84, 85) and AT1 angiotensin receptor (52).

For several recycled proteins, including transferrin receptor, two alternative recycling pathways have been described: a fast direct route from early endosomes or earlier compartments to the plasma membrane and a slower route involving the transport of cargo proteins to the endocytic recycling compartment (ERC) before transport to the plasma membrane (49, 86). The detection of recycled AIP56 after a 2-min chase suggests that part of the toxin follows the fast recycling route, while the observed colocalization with Alexa 647-transferrin in a perinuclear compartment at later stages suggests that a pool of the toxin is recycled from the ERC (slow recycling). Based on the observation that AIP56 recycling requires PI3K activity, we speculate that the fast recycling route plays a major role, since it is known that PI3K is mainly involved in the fast recycling (52, 54, 85), with very little evidence of its involvement in slow recycling (87).

The efficient sorting and recycling of several cargos/receptors internalized by receptor-mediated endocytosis require exposure of ligand-receptor complexes to low endosomal pH (52, 60). Frequently, this leads to the dissociation of the ligands from the receptors (52, 60). The membrane-bound receptors are recycled, while the released ligands concentrate in the vesicular parts of the early endosomes and are then routed through a microtubule-dependent maturation process to late endosomes/lysosomes (49). Our results suggest that endosomal acidification is dispensable for the recycling of AIP56, because the appearance of recycled toxin into the extracellular medium was not affected by inhibition of the vacuolar ATPase pump with ConcA.

Although the release of endocytosed toxins back into the extracellular medium has been previously described for the plant toxin ricin (88–90) and for anthrax toxin (91), neither of those toxins has been reported to follow the endocytic recycling route (91, 92). The release of endocytosed ricin seems to occur through a mechanism involving exocytosis of multivesicular bodies (89), allowing intoxication of bystander cell populations (90). In the case of anthrax, it has been demonstrated that the toxin is delivered into the extracellular medium in exosomes, contributing to prolonged toxicity and possibly to transmission of the toxin to distant cells (91). During this work, several attempts were made to detect activity of recycled AIP56. These consisted of large-scale recycling assays in which we recovered 48 ml of extracellular medium containing recycled AIP56 from eight 24-well plates and tried to concentrate the recycled toxin using either centrifugal devices or an immunoprecipitation protocol. However, we were unable to detect recycled toxin activity. This was likely due to significant losses of toxin during the procedures, since Western blotting of the concentrates revealed that the amount of AIP56 recovered was much less than expected and insufficient to perform activity tests. Therefore, the relevance of the putative recycling pathway and of recycled AIP56 for pathogenesis remains to be investigated.

Although monitoring the intracellular localization of fluorescence-labeled AIP56 by microscopy allowed imaging of some of the details of its intracellular trafficking, the low sensitivity restricts its usefulness for detecting translocation of the toxin into the cytosol. Nevertheless, several biochemical findings indicate that endosome acidification and low-pH-triggered conformational changes are involved in AIP56 translocation: (i) the arrival of AIP56 at the cytosol requires endosome acidification, since AIP56-dependent p65 cleavage was prevented by inhibitors of the vacuolar ATPase pump (61); (ii) at pH 5.5 or lower, AIP56 becomes more hydrophobic and is able to interact with artificial lipid bilayer membranes; (iii) a low-pH pulse promotes translocation of cell surface-bound AIP56 across the cytoplasmic membrane into the cytosol, similar to what has been reported for several toxins that translocate from endosomes through a pH-dependent mechanism (31, 32, 68). Since AIP56 reaches early endosomes and then follows the recycling route, translocation into the cytosol likely occurs from early and/or recycling endosomes. Early endosomes are the first acidic compartment encountered by the toxin upon endocytosis, and therefore, it is probable that at least the beginning of the translocation process occurs in this compartment. The fact that in less than 30 min, when most of the toxin is still located in EEA-1-positive vesicles, an amount of AIP56 sufficient to intoxicate the cells is already at the cytosolic compartment together with the fact that inhibition of PI3K activity does not inhibit AIP56 toxicity fully agrees with the hypothesis that early endosomes are the first organelle along the recycling route from which the toxin translocates into the cytosol. The observation that microtubule depolymerization does not affect AIP56-mediated p65 cleavage is compatible with this interpretation and confirms that late endosomes/lysosomes are not involved in AIP56 intoxication, as it is known that maturation of early endosomes into late endosomes is a microtubule dependent process (93).

In conclusion, with this study we elucidated important aspects of AIP56 trafficking, and we show that these are conserved between fish and mammalian cells. Our findings reveal that AIP56 is a short-trip toxin that follows two intracellular pathways that are at least partially independent: the intoxication pathway (that requires acidification of the endosomes and involves conformational changes in the toxin molecule but is independent of PI3K activity) and the recycling pathway (independent of endosome acidification but requiring PI3K activity). The evidence that a significant pool of endocytosed AIP56 follows the recycling pathway and is released into the extracellular medium suggests a novel intracellular route among the ones already established for bacterial AB toxins. Additional studies are required to determine whether AIP56 recycling has any relevance for P. damselae subsp. piscicida pathogenesis.